The education systems of Latin America show poor results in terms of providing quality of education equitably. In part this is because the region spends less per student at all levels than most regions in the world. This gap in per pupil expenditures between Latin America and the world is growing.

Although relative to overall government expenditure, countries in Latin America spend as much in education as do other regions, relative to gross national product they spend less. This suggests that the underfinancing of the sector is a result of the low tax base rather than of low government priority to the education sector. The ratio of expenditures per students to per capita GNP between different levels is smaller in Latin America than in Africa and is similar to that in Asia. In Latin America a higher proportion of education budgets is spent in recurrent expenditures, and a higher share for higher education. The share of enrollments in the private sector is at the high end in Latin America, relative to other regions.

The paper concludes by proposing options for reform including, but not limited to, increasing the level of resources for education. Policy analysis capacity should be developed in Ministries of Education to utilize existing resources better. Resources should also be used to promote innovations to improve the mix of inputs and the technical efficiency of education.

This paper examines the financing of public education in Latin America, and compares the patterns of financing in the region with those of other regions in the world. In conclusion, the paper proposes options to address the constraints identified in the analysis.

The analysis is based on original computations using data from UNESCO's World Education Report 1991. For comparative purposes countries were grouped in eight geographical regions: Sub-Saharan Africa, Other African countries, Latin America, U.S. and Canada, Other countries in the Americas (including the Caribbean, Belize, Suriname, Guyana), Asia, Europe and the former Soviet Republics and Oceania. Appendix 1 lists the countries which were included in each region.

The paper is structured in ten sections. The first section of the paper outlines the problem of low levels of performance of the education systems and links that problem to the theme of finance.

Section II examines the absolute levels of expenditure in education for primary, secondary and tertiary education for Latin America and other regions of the world. Section III analyzes the growing gap in education expenditures between Latin America and other regions. Section IV discusses the level of effort government makes to finance education, which leads to an analysis of the level of effort in education relative to national resources (section V), and to an examination of the effort to fund students at different levels (section VI). Section VII discusses the relative emphasis in each level of education, analyzing per pupil expenditure at each level relative to other levels. Section VIII reviews the distribution of education expenditures by type and level. Finally (section IX) the paper discusses the role of privately financed education at different levels. The last section of the paper discusses the findings of this study and proposes recommendations to address the major financial constraints in the context of the undergoing economic transformations in Latin America.

In each section I first examine how all countries in the world, for which data are available, perform on each indicator, and then examine the performance of Latin America relative to other regions. In addition, I also provide comparative evidence from countries in other regions, which have levels of per capita income that fall within the same range as the levels of income per capita in Latin America.1

Latin America has the doubtful distinction of being the region with the most inequitable income distribution in the world.2 As leaders of the countries of the region recognize the need to systemically reduce inequities, increase competitiveness, consolidate democratic institutions and promote stability and accountability, they will increasingly have to turn to educational institutions.

A look at the education systems of Latin America, however, shows systems that are turning out products of dubious quality and at low efficiency. Children are completing the first four grades of primary school without basic reading and math skills. Secondary school graduates seem to have been better prepared to continue onto higher education than to find opportunities for productive employment. Graduates at the tertiary level abound in fields that are at some distance from the most productive or dynamic sectors of the economy. In sum, the education system at all levels seems to be doing a better job at providing credentials than at developing basic skills, entrepreneurial abilities, critical thinking, citizenship, talent or social responsibility.

A recent study of the International Association of Educational Achievement assessed the reading ability of children in school at ages 9 and 14. Using equivalent instruments students were tested in 32 countries. Venezuela was the only Latin American country included in the study.3 The results show that Venezuelan students have one of the lowest levels of reading ability in the world. Fourth grade students have on average the lowest level of achievement of all 32 countries studied. In the section of the test measuring the ability to obtain information from tables or instructions, the average Venezuelan student has a lower ability than the 5% lowest performing students in Finland, Hong Kong or the United States, the countries with highest scores. The best 5% among the Venezuelan students have lower results on the test than the average student in Finland, Hong Kong or the United States. In grade 9, the performance of Venezuelan students is one of the four worst in the 32 countries, followed by Nigeria, Zimbabwe and Botswana.

A standardized test in mathematics and science administered to students in Brazil, China, England and Portugal shows that Brazilian students have the lowest scores. While Chinese students scored 80% of the answers in the math test correctly, students in Sao Paulo only scored 37%, and those in Fortaleza only 32%. In science Chinese students obtained 67% of the answers on the test correctly, but students in Sao Paulo scored only 53% and those in Fortaleza only 46%.4

The problem of low quality educational results is compounded as students progress through the system. A recent education assessment in Paraguay identified that many entrants to university had serious limitations in reading and writing ability.5

In addition to generating low levels of learning, schools also fail to reduce social inequities, as the opportunity to learn is smaller for rural and poor children. Data from Chile obtained administering a basic skills test to primary school students show that while students from families in the highest quintile of the income distribution answer 80% of the questions correctly, students in the lowest quintile answer only 40% of them correctly.6 A test administered to 3,248 primary school students (in grade six) in a random sample of Mexican schools shows that on average they answered only 48% of the items correctly in a curriculum test of basic subjects.7 Students in Mexican private schools obtain higher scores (65%) than their counterparts in public schools (47%).

There is little disagreement that the product of schools leaves much to be desired. A related problem is that schools turn out their product in very inefficient ways. Grade repetition (estimated at 40% for first grade and 30% for primary education, at a cost of almost 2 billion dollars per year)8 is a serious constraint on the internal efficiency of the system and another sign of low quality.

The educational establishment also perpetuates, rather than helping reduce, the inequalities that characterize Latin American societies. Although most of the children of school age in the region are enrolled in primary school at some point in their lives, there are great disparities between and within countries in the region in how many children are left out of school. Furthermore, for those who are enrolled there are great disparities in the opportunity to learn for children from different socioeconomic backgrounds and for those living in urban and rural areas. As a result, although many enroll in primary school at some point in their lives, many enroll in schools that put them at such disadvantage--compounded by their own social disadvantage--that school failure is the most probable outcome; many of these children learn little, repeat several times, and eventually drop out of school.

The children who do not have access to education are disproportionately children from poor and rural families. In Costa Rica, for instance, according to survey data from 1982, 13.3% of the children in the age group 7-12 with no schooling were from the lowest income quintile and only 2.6% from the two highest income groups.9 In Bolivia, where only 80% of the six-year olds and 93% of the seven-year olds in the bottom quintile are enrolled in school, the corresponding figures for the highest quintile are 95% and 98% respectively. At age seven most children in the highest quintile are enrolled in school, while for the bottom quintile this only happens at age nine. While 94% of the 15-year olds in the highest quintile are still enrolled in school, for the bottom quintile enrollment drops below 90% after children reach 13 years of age. (Figures derived from Encuesta de Hogares 1990, Bolivia. This survey was administered in the major cities in Bolivia in 1990. Only households located in cities with more than 10,000 persons were included in the sample.)

Table 1 shows that in El Salvador, children from poor homes are less likely to enroll in school, do it later and stay fewer years in school than their wealthier counterparts. While in the poorest 20% of the population only one in every two children are enrolled in school by age seven, in the 10% wealthiest group nine out of ten children attend school at that age. For the poorest 20% of the population only three out of every four children attend school at age nine for three years, while for the wealthiest 10%, three out of four children attend school by age five, nine out of ten attend by age seven and until age 14.

TABLE 1

A related fact constraining access is the different quality of the services provided to children from different groups. With regard to public education, as with many other fields of State activity in Latin America, those who have more, get more.

It should not be surprising that proportionately more people in rural areas have no access to schooling since those are the areas least served by the State in providing education. Many of the schools in rural areas have teachers teaching more than one grade (a rare phenomenon in urban schools), and have teachers with less training, supervision and access to materials. Many of the rural schools also do not offer all grades of primary education. In 1987 in Colombia 23% of the urban teachers were untrained versus 39% of the rural teachers; in Honduras the figures were 15% versus 46%; in Nicaragua they were 32% versus 74%.10

In Peru the percentage of trained teachers (maestros titulados) in primary education ranges from 95% in Arequipa (where the reported repetition rate is 11%), or over 70% in Lima (repetition rates around 10%), down to 20% of trained teachers in Madre de Dios (repetition rate 46%).11

Table 2 shows that the percentage of public "incomplete schools" (not offering all grades) is much higher in rural than in urban areas.

TABLE 2

Given these problems, what is the significance of educational finance? Education funds are necessary to pay for teachers' salaries, for the construction of schools, for teaching materials and other resources. The question of finance concerns how much should be spent to accomplish the task of teaching effectively, how resources should be allocated and who should pay.

The rest of this paper examines the patterns of response that the countries of Latin America have given to these three questions, and how those compare to the responses given by other regions. The paper concludes with a discussion of the significance of these patterns in the current context of the economic transformation being experienced by Latin America.

For all countries of the world expenditures per pupil in primary school (in 1989) range from US$ 11 to US$ 8,400. On average countries for which I have data spend US$ 887. Half of the countries spend US$ 162.

Latin America spends, on average, less per primary school student than any other region, with the exception of Africa. As table 2 shows, Latin America spends one fifth per primary school student of what is spent by Asia or the Caribbean, and less than one twentieth of what is spent by Europe or the U.S. and Canada. Countries of comparable per capita income in Asia spend, on average, 20% more per primary school student than countries in Latin America. Countries of similar income levels in the Caribbean spend almost five times as much as Latin American countries.

TABLE 3

Education spending per primary student in Latin America ranges from US$ 24 to US$ 255. On average, countries spend US$ 117; half of the countries spend less than US$ 96. Two in five countries in the world spend more than the highest spending country in Latin America at this level.

Expenditures on secondary education range from US$ 19 per pupil to US$ 6,712. On average, countries spend US$ 1,013 per secondary school student. Half of the countries spend less than US$ 268.

Latin America also spends less, on average, per secondary school student than any other region in the world as seen in table 4. African countries spend twice as much per student at this level, the Caribbean spends three times as much, and Asia spends five times as much. Europe, the U.S. and Canada spend 19 and 34 times as much, respectively. Latin America spends less per secondary school student than countries in other regions of comparable levels of income per capita.

TABLE 4

Spending on secondary school students in Latin America ranges from US$ 32 to US$ 376. On average, countries in the region spend US$ 167 in this level; half of the countries spend less than US$ 104 per student. Two in five countries in the world spend more than the highest-spending country in Latin America per secondary school student.

Expenditures per student in higher education range from US$ 33 to US$ 13,536. On average, countries spend US$ 3,079 per student at this level. Half of the countries for which I have data spend less than US$ 1,924.

At the tertiary level Latin America spends, on average, less per student than any other region in the world as seen in table 5. Countries in Sub-Saharan Africa spend three times as much per student at this level; countries in Asia spend four times as much. The U.S. and Canada spend 14 times as much at this level. Asian countries of comparable income levels spend 50% more per student in higher education than countries in Latin America.

TABLE 5

Spending per student in higher education in Latin America ranges from US$ 33 to US$ 1,709. On average, countries spend US$ 649; half of the countries spend less than US$ 457 per student at this level. Fifty-four percent of the countries in the world spend more than the highest spending country in Latin America.

Between 1980 and 1988 education expenditures in real terms decreased in 28% of the countries. Changes in education expenditures ranged from decreases of 14.7% per year to 25% average annual increases. In half of the countries the increase was less than 2.9% per year. On average, expenditures grew 2.9% per year.

The rate of growth of educational expenditures in real terms during the 1980s was slower, on average, in Latin America than in any other region of the world. Education expenditures in real terms increased 3 times faster in Sub-Saharan Africa, 3 times faster in Europe and the former Soviet Republics, 5 times faster in the U.S. and Canada, and 11 times faster in Asia. These results are summarized in table 6.

TABLE 6

As an indicator of government support for education, I examined the percentage of education expenditures relative to all public expenditures. This indicator suggests that the governments of Latin America make efforts that are comparable, in relative terms, to those of governments in other regions. This is paradoxical given the fact that we have seen how, in absolute terms, spending per pupil in Latin America is smaller than in any other region. This paradox is resolved in the following section which shows how education spending relative to GNP is lower for Latin America than for any other region. This suggests that the gap in education spending between Latin America and the world is a function of the low tax base in the region.

In 1988 education expenditures ranged from 4.7% of public expenditures to 27%. Half of the countries in the world spent less than 14.5% of the public budget on education; on average they spent 14.8%. Table 6 shows that Latin America is at the high end of education spending relative to public expenditures.

TABLE 7

A pervasive problem in the region is the gap between allocated and executed budgets. For example, in 1991 the Ministry of Education of Guatemala spent only 83% of the allocated budget for recurrent expenditures, and only 49% of the budget for capital expenditures.12 This reflects a common problem: the lack of financial expertise in the Ministries of Education that prevents fluid communication and negotiation with Ministries of Finance. As Ministries of Finance have implemented more mechanisms to discourage spending (as a response to the adjustment process), Ministries of Education have been unable to disburse the funds allocated to finance their activities (particularly funds which are not for salaries). This in turn reduces the prospects of negotiating larger budgets for these purposes in future years.

Even though the level of governments' relative commitment to education is similar in Latin America to that of other regions, education expenditures are substantially lower relative to per capita income because Latin American governments spend less than other regions relative to gross national product (GNP).

In 1988 expenditures in education ranged from 1.4% of GNP to 10.1%. Half of the countries in the world spent 4.7% of their national product on education; on average they spent 4.9%.

In 1988, on average, Latin America spent less on education relative to GNP than any other region in the world, as seen in table 8. Countries in Latin America spend, on average, half of what their counterparts in the Caribbean, or the U.S. and Canada, spend in education relative to GNP. The same pattern is observed if we restrict the comparison to countries within the same range of per capita GNP.

TABLE 8

One in three countries in Latin America is included in the lowest half--in terms of spending on education as a percentage of GNP--of all countries in the world. The exceptions are Cuba (6.7%), Honduras (4.8%), Nicaragua (6.2%) and Panama (5.6%). Eighteen percent of the countries in the world spend more than the highest spending country in Latin America (Cuba with 6.7%) and 20% of the countries spend more than the next highest spending country (Nicaragua with 6.2%).

Latin America spends less per pupil relative to per capita income than any other region of the world at all levels.

In 1988 countries' per pupil spending in primary school ranged from 3% of per capita GNP to 91% of per capita GNP. Half the countries spent 12% or less of GNP per capita per primary school student. On average, countries spent 14% of per capita GNP per primary school student.

Expenditures per primary student relative to per capita GNP are lower in Latin America than in any other region in the world. Among countries within the same range of income per capita, Asian countries spend 19% more than Latin American countries, countries in Sub-Saharan Africa spend 63% more, and countries in the Caribbean spend 3.5 times more.

TABLE 9

In 1988 countries spent from 3% of per capita GNP to 1.7 times per capita GNP on every secondary school student. Half of the countries spent 22% or less of per capita GNP per secondary school student. On average, countries spend 30% of per capita GNP per secondary school student. Expenditure per pupil in secondary school relative to per capita GNP is lower in Latin America than in any other region of the world, including countries within the same range of per capita income, as seen in table 10.

TABLE 10

In 1988 expenditures per university student ranged from 2% of per capita GNP to 16 times per capita GNP. Half of the countries spent 63% or more of per capita GNP per university student. On average, countries spend 1.8 times per capita GNP per university student.

Expenditures per university student relative to per capita GNP in Latin America are higher than in the U.S. and Europe as seen in table 11.

TABLE 11

For countries of comparable per capita income, Latin America maintains a relative balance between expenditures in university students and in students in primary and secondary education, and between spending in secondary and primary school students. The gap between spending in these levels, by contrast, is much larger in Africa.

Most countries spend more resources per university student than per student in primary education. Countries range from those which spend 11% of expenditures per university student than per primary student (only four countries spend less for university students than primary students) to those who spend 123 times more per university student than per primary school student. Half of the countries spend six times or more per university student than per primary school student. On average, countries spend 15 times as much per university student as per primary school student.

On average, Latin America spends substantially less on university students (table 12) relative to primary school students than Africa and less than countries in the Caribbean. This ratio is slightly higher than that for Asia, but three and four times higher than the respective ratio in the U.S., Canada and Europe.

TABLE 12

Most countries also spend more on each university student than on each secondary school student (only four countries spend more on secondary school students than on university students). Countries range from those who spend 10% of what they spend on each secondary school student on university students, to those who spend 25 times more on each secondary school student than on each university student. Half of the countries spend at least four times as much on university students as on secondary school students. On average, countries spend six times as much on university students as on secondary school students.

Latin American countries spend, on average, five times as much on university students than on students in secondary school. This is smaller than the respective ratios in Africa and the Caribbean and almost the same as the relative figure in Asia. Expenditures on university students relative to secondary school students are higher than in the U.S., Canada and Europe (table 13).

TABLE 13

Per pupil expenditures in secondary school students are generally higher than per pupil expenditures on primary school students. Countries range from those that spend 18% of per primary pupil expenditures on secondary school students to those who spend 25 times as much on secondary students as on primary school students. Half of the countries spend at least 1.8 times as much per secondary school students as per primary school student. On average, countries spend 2.6 times as much per secondary school student as per primary school student.

On average, Latin American countries spend 50% more on every secondary school student than on every primary school student. This difference is substantially lower than in Africa, and very close to the same ratio in the Caribbean, the U.S. and Canada, Asia and Europe (table 14).

TABLE 14

Education funds are used to finance different types of inputs. At one level they can be used to finance investments or recurrent expenditures, mostly teacher salaries. They can also be allocated differentially to various levels of education.

Latin America spends a substantially higher proportion of education budgets for recurrent expenditures than other regions. This suggests that there are relatively fewer public resources available for school buildings and other school investments. Latin America spends a slightly smaller proportion of the recurrent budget on teacher salaries than other regions, which suggests more resources available for administration and classroom materials.

The shares allocated to primary and secondary education are lower in Latin America than in the rest of the world, while the share for higher education is higher. This, compounded by the low overall level of education spending, explains the large gap in absolute terms between Latin America and the world at the lower levels of education.

In 1988 recurrent expenditures as a percentage of education expenditures ranged from 64% to 100%. Half of the countries spent over 92% of their education budgets on recurrent expenditures; on average they spent 90% on recurrent expenditures.

Latin America is at the high end of allocations to recurrent expenditures relative to the total education budget as seen in table 15.

TABLE 15

In the world, 50% of the countries spend 10% or more of their education budget on nonrecurrent expenditures, but in Latin America 93% of the countries spend 10% or less for this purpose. Forty-six percent of Latin American countries spend less than 5% of their education budgets on non-recurrent expenditures.

In 1988 the share of recurrent expenditures for teacher salaries ranged from 35% to 96%. Half of the countries spent 65% or more on teacher salaries. On average, countries spent 66% of their recurrent budget on teacher salaries. Latin America does not spend a higher share on teacher salaries than other regions in the world, as seen in table 16.

TABLE 16

In 1988 the share of recurrent expenditures for primary education ranged from 18% to 94%. Half of the countries spent 44% or less on primary education. On average, countries spent 46% on primary education.

The share of education expenditures for primary education in Latin America is lower than the equivalent share for other regions as seen in table 17.

TABLE 17

In 1988 the share of recurrent expenditures for secondary education ranged from 4.8% to 61%. Half of the countries spent 29% or less on secondary education. On average, countries spent 29% of their recurrent budgets on secondary education.

The share of recurrent expenditures for secondary education is also smaller in Latin America than in other regions, as seen in table 18.

TABLE 18

In 1988 the share of recurrent expenditures for higher education ranged from .2% to 40%. Half of the countries spent 17% or less of their recurrent budgets on higher education. On average, countries spent 17% of their recurrent budgets on higher education.

The share of recurrent expenditures for higher education is higher in Latin America than in Asia, Europe, Sub-Saharan Africa and the Caribbean.

TABLE 19

The share of students enrolled in private schools in Latin America is comparable or higher to the same figure in other regions.

In 1988 the percentage of pre-school enrollments in private centers ranged from 4% to 100%. Half of the countries had 42% or less of the enrollments at this level in private centers. On average, 49% of the children enrolled in pre-school attend private centers.

The share of pre-school enrollment in private schools is lower in Latin America than in all other regions except the United States and Canada.

TABLE 20

In 1988 the percentage of primary school students attending private institutions ranged from 0% to 100% in different countries. Half of the countries had less than 8% of primary school students enrolled in private schools. On average, 15% of primary school students are enrolled in private institutions.

The share of primary school children enrolled in private schools in Latin America is among the highest in the world, as seen in table 21.

TABLE 21

In 1988 the percentage of secondary school students enrolled in private school ranged from 1% to 82%. Half the countries have 16% or less of the secondary school students enrolled in private school. On average, 24% of the secondary school students attend private school.

The share of secondary school enrollments in Latin America is among the highest in the world, as seen in table 22.

TABLE 22

On the basis of comparisons with other regions, this paper concludes that Latin America has serious problems in the financing of education, in terms of imbalances between levels, but especially in terms of overall underfinancing of the sector. It is therefore not surprising that the education systems of the region are turning out products of low quality and at great inefficiency. It is clear that as the economies of Latin America open up to compete in global markets--where comparative advantage is increasingly a function of knowledge value added--attention to these constraints of the effectiveness and efficiency of the education systems becomes imperative.

While I do not propose that the solutions to the problems facing the education systems of Latin America are simply a matter of resources,13 resources should not be underestimated either. There is a point where attempting to improve starving education systems becomes like trying to squeeze water out of stones. Furthermore, overcoming the existing "system fatigue" and turning the gradual decline of the sector back into progress will require closing the gap between the level of resources for education in Latin America and those of other regions of comparable income levels.

A comparative perspective--though limited to the extent that data are not available for all countries of the world--is useful to place the patterns observed in the region in context. For instance, a recent World Bank report on education and human resources for Latin America states:

The key issue within the sector is the inefficient resource allocation between public spending on primary education on the one hand and tertiary education on the other. In most LAC countries, government subsidies still tend to favor tertiary education despite the fact that primary and pre-school education given higher returns per dollar or peso spent. For LAC as a whole, in 1989, higher education costs were subsidized at a rate seven times greater than those for primary education.14

However, the report fails to indicate that for countries with income levels falling within the range of income levels of Latin America, the same ratio of subsidies for tertiary to primary students is 26 in Sub-Saharan Africa, 14 in the Caribbean and 6 in Asia. The same report also fails to indicate that there has been a steady trend in Latin America to reduce those imbalances during the last 30 years. For instance, while the ratio of tertiary per pupil expenditure to primary per pupil expenditure was 16.6 in 1965, it declined to 13.6 in 1970, 11.5 in 1975 and

to 7.8 in 1978.15 Also ignored in the report is the fact that Latin America has achieved greater access to each level of education than most countries with comparable income levels. While there are problems that need to be addressed, this should be done without jeopardizing this comparative advantage.

The evidence examined supports the conclusion that Latin America invests less per student at all levels than most regions of the world and that this gap between Latin America and the rest of the world is growing.

The paradox of lower spending levels in absolute terms and relative to income levels in the region, while the level of effort of the governments is comparable to that of other regions, leads to the conclusion that one of the problems undermining the education sector is the low tax base with which governments operate.

A critical issue then is to achieve a significant increase in education resources at all levels. This will not be achieved by transfers of existing resources between levels, for it would further increase the gap in per pupil expenditure between Latin America and the rest of the world.16 Rather additional resources must be generated for education at all levels.

Options to increase resources for education include increasing the share of education in the government budget. This would mean increasing the level of effort of government for the sector beyond that of other sectors. The critical question here is what should be sacrificed for education's sake. To the extent that education gains resources currently spent in less productive activities, such as defense or internal security, for instance, the costs of this adjustment will be lesser than if resources were transferred from other important sectors such as agriculture, health, housing or transportation.

Another option is to maintain the level of education relative to total government expenditures, increasing the tax base and the overall levels of spending of the government. This option would be inconsistent with the current trend of "less government."

A mixed option would be increasing government resources, but only for education, which would require broadening the tax base and targeting the additional resources for education. This trend is consistent with proposals to reduce the scope of activities where the government intervenes, but to concentrate government action in areas with externalities and where the government has a comparative advantage, such as education, health, construction of roads, etc.17

Other options for expanding resources include promoting private contributions, either in the form of additional financing from beneficiaries of education (tuition fees), or promoting the expansion of private schools. As will be discussed later, neither of these options is an appropriate response to generate additional resources at the basic levels in the current context of Latin America.

An option to generate additional resources for a particular level or type of education is to transfer resources from other levels. Given the underfinancing of expenditures at all levels, this option is not considered an appropriate response to the problems facing the education system, nor is it considered more feasible than transferring funds, for example, out of the Ministry of Defense and into the Ministry of Education.

The distribution of public education expenditures by type is not particularly different from those of other regions. However, Latin America spends a higher percentage of public education funds for tertiary education than other regions, to the detriment of other levels of education. While this reflects the greater effort made in Latin America to expand educational opportunity at the tertiary level, it also suggests inequitable use of public funds, since students at this level have a higher capacity to contribute to the cost of their education than students at lower levels of education.

The shares going to different levels of education should be changed in favor of basic and secondary education, but this should be done in the context of a growing education budget, and not draining resources from universities. A recent study of the World Bank is consistent with this assessment stating that per pupil expenditure in higher education in Latin America is not excessive:

Higher education in Latin America has gone through difficult fiscal times. Real expenditures per pupil have decreased, non-personnel outlays have diminished, and, by inference, undergraduate instructional quality has declined. Strong arguments can be given that per pupil expenditures should be increased, but existing misallocation of resources provides no assurance that additional spending would be used efficiently.18

The existing evidence offers no reason to expect significant gains in resources from further privatization of education, a frequently touted option for the sector. Latin America is at the high end of enrollments in private institutions at all levels. The development of private schools could be promoted by reducing the negative incentives that currently stem from excessive government regulation (e.g., caps on tuition fees, excessive paper work and "permisology," which gives supervisors of the private school system the ability to harass private schools, particularly small government non-elite schools that lack the political clout to protect themselves).

Promoting privatization, however, is a good option which favors more efficient utilization of existing resources rather than generating additional resources. Channelling public funds to privately managed schools serving poor children (such as the Fe y Alegría program) or establishing matching funds to support schools serving poor children are options to promote privatization.

The option of generating additional resources at the basic levels by charging tuition fees in public schools should be carefully considered. Most public schools already have some forms of "voluntary" contributions from parents. Further cost-recovery at primary and secondary education levels should not be introduced unless appropriate studies of price elasticity of demand are made to anticipate the impact on equity and the expected gains in levels of resources. It may be desirable to increase cost-recovery at the tertiary level, with mechanisms that ensure that this does not discourage enrollments from needy students, but this is more justifiable as a strategy to increase the internal efficiency of the sub-sector by placing incentives to encourage early completion rather than as a substitute for public expenditures on the sector.

The impact of financial constraints is compounded by the fast pace with which Ministries of Education have had to adjust to rapid changes during the last decade. The economic crisis that affected most countries of the region during the 1980s had an important toll on the education sector.

This toll, and the "system fatigue" it generated in Ministries of Education, is a constraint to options to improve the use of existing resources. In the short run additional resources will be necessary to fuel reforms designed to make better use of resources. But at the same time, additional resources should not be "thrown" at fatigued systems. Fresh resources should be used as opportunities to promote improvements in quality and internal efficiency, not to promote more of the same. For instance, it is quite possible that in systems with large rates of repetition, quality improvements in the forms of textbooks, instructional supplies and better teacher training over time will pay for themselves.19 However, developing these new instructional materials and providing them in the short run means additional up front investment before any savings from reduced repetition have been realized.

One of the priorities to achieve more resources for education and better utilization of existing resources is to develop the capacity in Ministries of Education for policy analysis, program preparation and budgeting and financial control. Many Ministries of Education lack the ability to establish priorities for action and to develop appropriate programs and budgets that can persuade Ministries of Finance to fund them. In addition, Ministries of Education have institutional constraints to disburse funds allocated to them or to respond effectively to the spending disincentives implemented by Ministries of Finance. As a result, it is hard for the sector to justify increases in allocations for nonsalary activities when assigned budgets are not spent.

Some of the options to improve internal efficiency include exploring alternative configurations of education inputs (e.g., more textbooks and less-trained teachers, teachers with less specialized training, more parental participation). Changing the mix of inputs may also reduce costs, but this should be done with the explicit goal of increasing efficiency rather than just cutting costs (e.g., developing new teaching technologies relying more on peer teaching and instructional materials, which allow increasing class size, or developing new teaching career structures to give more incentives to skill improvement rather than to credentials, seniority or political connections, promoting new forms of local management, transferring some of the tasks now handled by large and inefficient central bureaucracies to more decentralized levels).

Other options include increasing the efficiency in the use of existing inputs, such as giving priority to the assignment of new teachers to understaffed schools, or establishing systems to encourage the better trained teachers to attend the early grades where school failure is higher, or channeling instructional resources to the early grades of primary school to reduce repetition rates.

Significant gains in the efficiency of higher education could be achieved by changing the formulae used for the allocation of public funds between universities. Typically, existing allocation practices reflect political weight and institutional tradition rather than the quality of graduates or the number of years it takes different institutions to produce a graduate. New criteria could provide incentives to encourage research, quality education, closer ties between programs and demands of the market, and on-time completion. University governance is another area that could benefit from innovations to move away from the deep-rooted systems of patron-client relationships and from the control of universities by strong political and interest groups toward more accountability and professional management.

Gains in efficiency and quality at the basic levels could be attained by developing combined schemes of publicly funded, but privately managed educational alternatives. An example is the Fe y Alegría schools which operate in 12 countries of the region and reputedly do a much better job than the State at providing quality education to the poor.20

Strengthening the ties between schools and communities is an attractive option, not to generate additional resources--which may not be feasible given the scenario described later--but to improve the utilization of existing resources, to establish accountability from teachers to the community and to monitor teacher attendance and performance.21

The 1980s were a decade of growing austerity in Latin America which reduced the potential contributions of households to education. Adjusting countries also reduced public education expenditures. Facing reductions in funds, education ministries disproportionately cut the share for teaching materials. Eventually, teacher salaries deteriorated in real terms too.

The announcement in 1982 that Mexico could not continue servicing its foreign debt obligations marked a rupture with the environment-facing governments in the Third World. As it became clear that many countries had been borrowing (and banks lending) too much, commercial banks became more reluctant to lend money. Faced with mounting pressure from interest payments, countries turned to multilateral banks who were willing to lend them more funds to support efforts to stabilize and adjust their economies. For many nations, economic adjustment was the hallmark of the 1980s.

Adjustment programs have had a relatively short-term focus, mostly because adjustment was initially thought to be a short-term problem. A UNICEF study for Latin America points out:

The experience with adjustment programmes, however, has not been satisfactory. While these programmes should not be seen as the cause of the economic decline of the 1980s, they clearly have not been able to reverse the adverse developments in the living standards of the poor, nor in most cases have they led to resumed economic growth. In addition, these adjustment programmes have generally made no explicit efforts to prevent deteriorations in human conditions. Thus there is still considerable debate on the possibility of continuing progress in health, nutrition and education under these adverse economic conditions and on the nature of the macroeconomic policies and health measures needed to foster such progress.22

There are two basic channels through which adjustment can influence education.

The first is through the impact of adjustment on households. As policies succeed at contracting aggregate demand, the living conditions of many people deteriorate, thereby reducing the income they have available for school supplies (uniforms, tuition, textbooks, etc.) and increasing the need for children to contribute to the household economy.

This type of impact would influence external and internal efficiency and equity. Poorer families would have to pull their children out of school. The reduction in the ability of households to contribute could also lead to internal inefficiencies not only from the reductions in inputs available to learning, but also because of diminished efficiency in the mix of inputs. For instance the effectiveness of teacher instructional practices may be reduced if students have no textbooks or notebooks.

A second mechanism linking economic adjustment and education provision is the change in public education finance that results from pressures to reduce public spending. Overall cuts in public spending may lead to disproportionate cuts in education expenditures vis-a-vis other government expenditures.

The short-term structural rigidities in the education budget make it easier to cut certain items such as funds for teaching materials or school repairs. This has a negative impact on the internal efficiency of education as it leads to inefficient changes in the mix of inputs. It is difficult for teachers to compensate for deteriorating buildings, or to preserve educational quality when there is a lack of textbooks and materials.

There are also institutional rigidities that can lead to reductions in total resources for education, increasing inequities in educational spending. For instance, urban students and schools are typically more vocal, better organized and closer to the distribution centers of the Ministry of Education. Hence, reductions in the available instructional supplies are likely to be faced with increased pressures from those parents to preserve their initial allotments, thus increasing their share of a declining total pie. The implication is that rural schools, or those attended by students whose parents have less political leverage, will have to face disproportionate reductions in their supply of chalk and other learning resources.23 Similarly, universities have more leverage (political and technical) to increase their share in a shrinking pie of total resources for education.

The 1980s were a decade of growing austerity in Latin America as per capita income declined in 78% of the countries. As a result, living conditions deteriorated in many countries. This is important as it effectively reduced the capacity of households to support the work of the school. The implication is that in a scenario of austerity-facing households the efforts of the State to maintain the same levels of education provision should be more, not less.

In highly unequitable economies, declining income per capita may lead to further inequities. For Latin America, some authors have proposed that a series of negative multipliers would translate a 5 to 10% reduction in GNP into a reduction three to four times larger for low income families. This notion stems from "the fact that minimum wages drop faster than do average wages, that the prices of essentials are subject to greater increases than is the Consumer Price Index, that the newly unemployed frequently also suffer the loss of health coverage and that cuts in public expenditures are typically asymmetrical. In terms of child welfare, these biases are further aggravated by the fact that poorer families generally have a larger than average number of children."24

In Latin America real salaries in manufacturing declined 8.4% between 1980 and 1985. Salaries in construction declined 19%, the minimum salary declined 11% and informal sector earnings declined 27% (Ibid, 19). Between 1981 and 1988 urban minimum wages in real terms declined 33% in Brazil, 46% in Mexico, and 40% in Peru.

All of these changes increase the marginal contribution of the work of children to poor households (even if their work will generate less income because of reduced demand) and diminish the ability of parents to further contribute to education by paying for school supplies, uniforms and making in-kind or cash contributions to the school. The combined effect of these factors is to increase the pressure to pull children out of school and send them to work.

Facing reductions in funds, education ministries disproportionately cut the share for teaching materials. Eventually teacher salaries deteriorated in real terms too. This made the provision of education more difficult, not just because teachers saw class sizes increase, but because it changed the organization of the teaching process: teachers had less time to prepare classes, fewer resources for teaching, taught in buildings in increasing disrepair, and faced students tired from increasing household responsibilities, sometimes hungry and sick and with fewer things to bring from home, such as notebooks and textbooks, that could help them learn. To what extent was education still possible under these circumstances? This question is not rhetorical, for it is possible that declines of education quality beyond a minimum threshold would lead parents to reassess the relative benefits of sending their children to school.

It is in this context that the low levels of education finance in Latin America and the exploration of policy options should be considered, for this context gives the growing gap between levels of education finance in Latin America and the world crisis proportions. It will take leadership to recognize the danger in this crisis and to construct opportunities for change.

Algeria, Egypt, Libya, Morocco, Namibia, South Africa, Togo, Tunisia

Angola, Benin, Botswana, Burkina Faso, Burundi, Cameroon, Cape Verde, Central African Republic, Chad, Comoros, Congo, Cote D'Ivoire, Djibouti, Equatorial Guinea, Ethiopia, Gabon, Gambia, Ghana, Guinea, Guinea-Bissau, Kenya, Lesotho, Liberia,

Madagascar, Malawi, Mali, Mauritania, Mauritius, Mozambique, Niger, Nigeria, Rwanda, Sao Tome and Principe, Senegal, Seychelles, Sierra Leone, Somalia, Sudan, Swaziland, Tanzania, Uganda, Zaire, Zambia, Zimbabwe

Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, Cuba, Dominican Republic, Ecuador, El Salvador, Guatemala, Haiti, Honduras, Mexico, Nicaragua, Panama, Paraguay, Peru, Uruguay, Venezuela

Antigua, Bahamas, Barbados, Belize, British Virgin Islands, Dominica, Grenada, Guyana, Jamaica, Netherlands, St. Kitts, St. Lucia, St. Vincent, Suriname, Trinidad

Canada, United States

Afghanistan, Bahrain, Bangladesh, Bhutan, Cambodia, China, Cyprus, Hong Kong, India, Indonesia, Iraq, Iran, Israel, Japan, Jordan, Korea Democratic, Korea Republic, Kuwait, Laos, Lebanon, Malaysia, Maldives, Mongolia, Myanmar, Nepal, Oman, Pakistan,

Palestine, Philippines, Qatar, Saudi Arabia, Singapore, Sri Lanka, Syria, Thailand, Turkey, United Arab Emirates, Vietnam, Yemen

Albania, Austria, Belgium, Bulgaria, Czechoslovakia, Denmark, Finland, France, Germany Democratic Republic, Germany Federal Republic, Greece, Hungary, Iceland, Ireland, Italy, Luxembourg, Malta, Monaco, Netherlands, Norway, Poland, Portugal, Romania, San Marino, Spain, Sweden, Switzerland, USSR, United Kingdom, Yugoslavia,

Australia, Fiji Kribati, New Zealand, Oceania, Papua New Guinea, Samoa, Tonga

* Fernando M. Reimers is an Institute associate at the Harvard Institute for International Development, where he specializes in education policy analysis, research, and planning. He also teaches a graduate course on education policy in Latin America at the Harvard Graduate School of Education. He obtained his doctorate in educational planning at Harvard University and has published several books and articles on education finance and development in Latin America. He has been a consultant for the Inter-American Development Bank, the World Bank, USAID, and UNESCO.

NOTES

1. The country with the lowest per capita income in Latin America in 1989 had US$ 400 per capita, and the highest US$ 2,620. Most tables include separate comparative information for countries in each region which fall within this range of per capita income. This excludes the poorest 25% and the richest 30% countries in the world.

2. World Bank, World Development Report 1990 (New York: Oxford University Press, 1990).

3. W. Elley, How in the World Do Students Read? (International Association for the Evaluation of Educational Achievement, 1992).

4. A. Lapointe et al., Learning Science, Report No. 22 (IAEP, 1992).

5. F. Reimers et al., Análisis del sistema educativo en el Paraguay. Sugerencias de política y estrategia para su reforma (Asunción: HIID-CPES, 1992).

6. E. Schiefelbein, "Financing Education for Democracy in Latin America" (Santiago, Chile: 1991). Mimeo.

7. G. Guevara, "México: Un país de reprobados", Nexos (June 1992): 33-44.

8. E. Schiefelbein, Repetition Rates: The Key Issue in South American Primary Education (Washington, DC: The World Bank, LATHR Division, 1988).

9. The World Bank, Costa Rica. Public Sector Social Spending (Washington, DC: The World Bank, 1990) 19.

10. Derived from table 22, page 53. UNESCO-OREALC, Situación educativa de América Latina y el Caribe. 1980-1987 (Santiago, Chile: UNESCO, 1990).

11. These figures are for 1985. T. Tovar, Ser maestro. Condiciones del trabajo docente en Perú (Lima, Peru: UNESCO-DESCO, 1989).

12. F. Reimers, Education Policy Priorities in Guatemala. An Agenda for Policy Dialogue (Guatemala City: 1993) 24.

13. For a discussion of a systemic agenda for education reform, involving the areas of finance, management and products of education, see F. Reimers, A New Scenario for Educational Planning and Management in Latin America (Paris: UNESCOInternational Institute for Educational Planning, 1990).

14. The World Bank, Human Resources in Latin America and the Caribbean (Washington, DC: The World Bank, 1993) 52-53.

15. E. Schiefelbein, Education Costs and Financing Policies in Latin America (Washington, DC: World Bank, 1987) 7.

16. It should be noted that countries of comparable income levels in Asia spend 20% more per primary school student, and in the Caribbean five times more per primary student than countries in Latin America. A similar gap exists for secondary students (where Asia spends 20% more, the Caribbean 2.4 times more and Sub-Saharan Africa 1.8 times more than Latin America) and for university students (where Asia spends 52% more per student, the Caribbean 3.4 times more per student, and Sub-Saharan Africa 3.64 times more than Latin America).

17. The World Bank, World Development Report 1991 (Washington, DC: Oxford University Press, 1991).

18. D. Winkler, Higher Education in Latin America. Issues of Efficiency and Equity (Washington, DC: World Bank Discussion Papers, 1990) 30.

19. This is, for instance, what was found in the Northeast of Brazil, see R. Harbison and E. Hanushek, Educational Performance of the Poor: Lessons from Rural Northeast Brazil (Oxford: Oxford University Press, 1992).

20. A study of this program in Guatemala found that 85% of the students who enter pre-school in Fe y Alegría complete six grades in seven years, in contrast with 34% completion rates in the same period in Government schools. See J. Diaz, La educación subvencionada. Un modelo de opción para la educación nacional (Guatemala: Universidad del Valle de Guatemala, Facultad de Educación, 1989) 67.

21. There are three educational innovations in Latin America that provide quality education to the poor where strengthened school-community relations play a critical role, see F. Reimers, Education and the Consolidation of Democracy in Latin America. Innovations to Provide Quality Education with Equity (Washington, DC: Academy for Educational Development, Project EHRTS, 1993).

22. T. Albanez, E. Bustelo, G. Cornia and E. Jepersen, Economic Decline and Child Survival: The Plight of Latin America in the Eighties (Florence: UNICEF, Innocenti Occasional Papers, 1989) 1.

23. For a study of how Ministries of Education responded to budget cuts see F. Reimers, "Tiene Jomtien relevancia en América Latina. Los ajustes a la educación cuando se ajusta la economía", La Educación 35.108-110 (1991): 101-132.

24. T. Albanez, E. Bustelo, G. Cornia and E. Jepersen, Economic Decline and Child Survival: The Plight of Latin America in the Eighties (Florence: UNICEF, Innocenti Occasional Papers, 1989) 17-18.